A few thoughts on viewport size when writing (typing):
When one composed on the Monotype or Lintotype hot metal casting machines, there was no way to preview what you have typed at all – no viewport. One could only find out after the type has been cast. In other words, writing blindly. (though composing on these machines can hardly be called ‘writing’, but it’s doable)
If you have a large viewport when composing, you can see where you are in relation to the stream of text that one has already written. One can navigate back when editing. Or the text can have some sort of semantic structure (tagged as you write) where you can navigate via an outline.
If you have a very tiny viewport that can only accomodate a few words at a time, then you can verify what you have input but still have no easy way to visualise what you have written, or navigate back. A one-character viewport would only be useful for verification.
What benefits would writing blindly be? Or there is a geographical separation between input (private) and output (public)?
(written with Markdown in Tumblr site)